Delineation of site-specific productivity zones using soil properties and topographic attributes with a fuzzy logic system

Abstract

A delineation procedure for site-specific productivity zones was developed with a fuzzy logic system using soil properties obtained from on-the-go electrical conductivity (EC) and organic matter (OM) sensors and topographic attributes. EC, OM, slope and curvature were used as input variables, and productivity was set as an output variable. The fuzzy rules were developed with grower's knowledge for typical central Kansas upland fields; areas within the field having high OM, low EC and low slope have the highest productivity potential, and areas within the field with low OM, high EC and high slope have the lowest productivity potential. The fuzzy logic system performed properly and generated productivity as designed by the fuzzy logic and inference scheme. To validate the system, an adjacent field with 5 years of wheat yield data was selected. The spatial agreement between productivity and yield showed as high as 0.57 and 0.35 for overall accuracy and kappa coefficient. The level of agreement is promising, considering there were many other yield-limiting factors such as precipitation, temperature and management effects. From comparison of the productivity map with the map generated by a fuzzy c-means clustering algorithm (FCM map), agreement between the productivity and yield exhibited generally higher in overall accuracy and Kappa coefficient than the agreement between FCM map and yield. Results of this study can benefit producers and consultants who utilise site-specific management by delineating productivity zones using EC, OM, slope and curvature from the on-the-go sensors.